In this chapter, we investigate the effectiveness of using a deep autoencoder network with three and five hidden layers. These networks will be used in combination with optimization algorithms to perform missing data estimation tasks. The results from these networks will be compared against those obtained from using the seven hidden-layered deep autoencoder network from the literature. The network training times are observed to increase with the increasing number of hidden layers.
CITATION STYLE
Leke, C. A., & Marwala, T. (2019). Deep Learning Framework Analysis. In Studies in Big Data (Vol. 48, pp. 147–171). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-01180-2_10
Mendeley helps you to discover research relevant for your work.